Department of Biostatistics Seminar/Workshop Series
Structual equation models in biomedical research: Uses and Abuses
Steven M. Brunwasser, Ph.D. Research Instructor in Obstetrics & Gynecology, Vanderbilt University School of Medicine
Structural equation modeling (SEM) provides a flexible framework for multivariate statistical modeling that may be highly useful in biomedical research. However, applied researchers often implement SEM techniques in suboptimal ways. This talk will provide a broad overview of SEM with a specific emphasis on biomedical research applications. The presenter will provide specific examples using biomedical data and highlight common challenges and missteps. The presentation will focus on basic SEM principles but will also touch on useful extensions and advanced topics (e.g., latent growth models, multiple-group models, multilevel SEM models, Bayesian SEM). Simple demonstrations will be provided using R (lavaan package) and Mplus software.